Bayesian Penalized-Likelihood reconstruction algorithm: noise reduction study of different contrast ratios on a 15 cm axial field-of-view PET/CT Scanner

The resolution and quantitative accuracy PET are significantly influenced by the reconstruction method. Bayesian Penalized-Likelihood reconstruction algorithm controls the noise amplification improving the image quality during the image reconstruction. The aim was to determine if the noise level can be reduced without a loss in image quality. The performance of BSREM was compared to ordered-subset expectation maximization (OSEM) for both phantoms (with different contrast ratios) and patient data acquired on a state-of-the- art PET/CT. The results confirm that for the phantom data, the number of counts can be reduced by a factor 2-4 using BSREM instead of OSEM. For the patient data, a similar trend was found. The possible count level reduction was at least a factor of 2. Noise reduction is possible by introducing regularization in the image reconstruction without a loss in image quality. This reduction can be used to lower the injected dose or shorten the acquisition time